124 NEW METHOD OF ESTIMATING NITRIC ACID, 
of saltpetre containing 10 grms. of refined sal tpetre in a de- 
militre; I then add, with a graduated burette, some drops of 
the solution of the protosulphate of iron, and agitate gently ; 
the mixture becomes brown, suddenly clears again, and ac- 
quires a more or less deep straw colour. I then continue to 
add in drops a normal solution, the strength of which I 
wished to ascertain, waiting each time until the liquid 
brightens; when these changes of colour become less rapid 
I apply a gentle heat ; and when they are no longer per- 
ceptible, I examine whether, on taking a few drops with a 
pipette, they still yield a blue colour with the red ferrocya- 
nide of potassium. When this blue colour appears, and no 
longer disappears on boiling the liquid, I note down the to- 
tal number of the divisions employed. A second examina- 
tion, in which a large number of the trials may be dispensed 
with, will yield an accurate result. 
Let N be the number of divisions of the normal solution 
requisite to saturate completely 25 cubic centimetres of the 
solution of refined nitre; I weigh off 10 grms. of the salt- 
petre to be examined, which I likewise dissolve in a demili- 
tre of water, and determine, in the same manner as was 
done for the solution of refined saltpetre, the number N 1 of 
divisions of the protosulphate of iron required to saturate 
completely 25 cubic centimetres of this solution ; and the 
fraction — gives the strength of the saltpetre under exami- 
tion with great accuracy. 
The process, which is applicable to the estimation of 
nitric acid, and of all. the nitrates in general, may be em- 
ployed for determining the quantity of nitric acid contained 
in gun-cotton. For this purpose, the whole of the cotton is 
oxidized with a mixture of sulphuric acid and peroxide of 
manganese ; the liquid is filtered, and decolorized with a 
few drops of protosulphate of iron, which decomposes the 
whole of the sesquioxide of manganese before acting upon 
